Hi all
I recently started using postgres more and using schemas in postgres to keep the number of databases under control.
When using schemas in a database and using rails, if two tables have the same name, the output of rake db:schema:dump is incorrect
Recreating a failing environment, have a database railstest with public, dev and test schemas:
$>rails test $>cd test $>vi database.yml development: adapter: postgresql database: railstest schema_search_path: dev encoding: utf8 username: railstest password: test host: localhost
# Warning: The database defined as 'test' will be erased and # re-generated from your development database when you run 'rake'. # Do not set this db to the same as development or production. test: adapter: postgresql database: railstest schema_search_path: test encoding: utf8 username: railstest password: test host: localhost production: adapter: postgresql database: railstest encoding: utf8 username: railstest password: test host: localhost $>rake db:sessions:create $>rake db:migrate $>rake db:schema:dump
The schema generated a this point is incorrect (look at the add_index lines): # This file is autogenerated. Instead of editing this file, please use the # migrations feature of ActiveRecord to incrementally modify your database, and # then regenerate this schema definition.
ActiveRecord::Schema.define(:version => 1) do
create_table "sessions", :force => true do |t| t.column "session_id", :string t.column "data", :text t.column "updated_at", :datetime end
add_index "sessions", ["session_id", "session_id"], :name => "index_sessions_on_session_id" add_index "sessions", ["updated_at", "updated_at"], :name => "index_sessions_on_updated_at"
end
I tracked this down to postgresqk_adapter.rb in the index method, it does not differenciate between schemas.
I have a small patch that corrects the problem for me, once the patch is applied, my schema.rb looks like : # This file is autogenerated. Instead of editing this file, please use the # migrations feature of ActiveRecord to incrementally modify your database, and # then regenerate this schema definition.
ActiveRecord::Schema.define(:version => 1) do
create_table "sessions", :force => true do |t| t.column "session_id", :string t.column "data", :text t.column "updated_at", :datetime end
add_index "sessions", ["session_id"], :name => "index_sessions_on_session_id" add_index "sessions", ["updated_at"], :name => "index_sessions_on_updated_at"
end
However I have no idea how to write a failing test for this ... 
The patch is attached (both fulltext and tar.gzipped in case) to this
email. I hope someone with more experience will pick it up and get it
in.
It applies cleanly against both activerecord-1.15.3/lib/active_record/connection_adapters/postgresql_adapter.rb and http://svn.rubyonrails.org/rails/trunk/activerecord/lib/active_record/connection_adapter/postgresql_adapter.rb rev 6891 (which is the last modified rev for this file as I am writing this email)
jean
postgresql_adapter.patch (1.11 KB)
postgresql_adapter.patch.tar.gz (584 Bytes)